How to check your suburb pages for duplication

You can check whether your suburb pages are too alike without any special software: take the text of each page, compare every pair, and see how much of the wording they share. We do this with a score called Jaccard overlap, and on four trade sites we measured in August 2026 the scores were between 0.46 and 0.57 before we touched anything. This post explains the measurement, shows what we found and describes the trap that waits after you fix it.

Why bother

Suburb or service area pages are the cheapest way to look local on a national map, which is why so many sites have a dozen of them. They are also the easiest pages to publish by copying one and changing the name. When we looked at what happened to a cluster of six such pages on a pest control site, covered in our case study on publishing six at once, visibility in the top 30 results fell within a week. A measurement you can repeat is better than a hunch about whether your own pages are at risk.

How the measurement works

1. Take the visible text of each page in the cluster, from the rendered page and not the source file. Menus, footers and lists of suburbs count, because a crawler compares them too.
2. Break each page's text into overlapping three word groups.
3. For any two pages, divide the number of groups they share by the total number of different groups across both. That is the Jaccard score.
4. Average the score across every pair in the cluster.

A script does this in a few lines. You can also do a rough version by hand: open two pages side by side, and count how many paragraphs you could swap between them without anyone noticing.

Report the unique word count per page alongside the score. Overlap on its own is only half the picture, as the next section shows.

What we measured on four sites

On 18 August 2026 we measured the location clusters on four of the trade sites we run, before and after removing the blocks every page shared:

The biggest single source of duplication was the review block, repeated word for word on every page. Taking it out alone moved the carpentry site's score from 0.418 to 0.201. After that came the identical four question FAQ, which also produced the same structured data on every page, then the short service descriptions in the service cards, and then two of the three body paragraphs.

The trap: fixed and thin

Look at the right-hand column. Every cluster got less alike, and every cluster also got a lot shorter in unique words. The carpentry site went from 323 unique words a page to 119. Removing what the pages shared exposed how little each page had of its own.

Under roughly 150 unique words we do not call a cluster fixed. It is converted from duplicated to thin, and the next job is to write real content for each page. On the builder site the picture was worse than the table suggests: the site has 28 area pages but local copy for only 10 suburbs, so 18 pages showed a one line generic introduction. Those pages never had anything to strip back to.

We still think the strip was worth doing. Our working assumption is that quality is judged across a whole site, so a duplicated cluster can drag down pages outside it, while thinness mostly hurts the thin pages themselves. That is our reading and not something Google has confirmed. But reporting only the drop in overlap would have made the result look better than it was.

Do not reach for noindex

On the builder site some of the thin pages with no local copy had higher impressions than the pages with full copy, at 265, 274 and 317. Hiding them would have thrown away the pages that Google was already showing. Check impressions per page before you decide which ones to remove or merge.

A repeatable check for your site

1. List every location page you have.
2. Score the cluster, or at least compare five pairs by eye.
3. Find the blocks that repeat. Reviews, FAQs and service cards are the usual carriers.
4. Count unique words per page after removing them.
5. For each page, write down one thing that is true only of that suburb. If you have nothing, merge that page into a wider area page.

Where this sits in our plans

Measuring a cluster, reading the pages behind it and deciding what to merge, rewrite or leave alone is the work of the SEO add-on. If your own location pages worry you, our free website audit is a starting point, and get started is how you ask us to look at them properly. Pricing is on the pricing page.